Cooperative lock override procedure

• US 6,757,769 B1
• Filed: 11/28/2000
• Issued: 06/29/2004
• Est. Priority Date: 11/28/2000
• Status: Expired due to Term

• Current Assignee: Emc IP Holding Company LLC (Dell Technologies Inc.)
• Original Assignee: EMC Corporation (Dell Technologies Inc.)
• Inventors: Ofer, Adi
• Primary Examiner(s): PORTKA, GARY J
• Application Number:
  US09/724,014
• Time in Patent Office: 1,309 Days
• Field of Search: 711/151, 711/152, 711/163, 711/147, 710/108, 710/200, 710/220, 710/240, 710/241, 710/244
• US Class Current: 710/200
• CPC Class Codes: G06F 11/0793 Remedial or corrective actions recovery from an exception in an instruction pipeline G06F9/3861; by retry G06F11/1402; for recovering from a failure of a protocol instance or entity H04L69/40
• CPC Class Codes: G06F 2209/522 Manager
• CPC Class Codes: G06F 2209/523 Mode
• CPC Class Codes: G06F 9/524 Deadlock detection or avoidance
• CPC Class Codes: G06F 9/526 Mutual exclusion algorithms

Abstract

Queued lock services for managing a shared resource in a data processing system include a cooperative lock override procedure. On detecting a protocol failure by another processor, the detecting processor confirms that the failing processor is the lockholder and passes the lock to the next requestor in the queue.

Claims

1. A method for providing cooperative queued locking and unlocking services for managing a shared resource in a data processing system including a plurality of processors as lock requestors, each processor supporting atomic operations and being coupled to the shared resource through one or more first common communication channels, including the steps of:

providing for each processor a monitoring procedure for detecting a predetermined indication of protocol failure by any one of the plurality of processors and identifying the failing processor;

providing for each processor a lock services procedure including a queuing procedure for unsuccessful lock requestors, locking and unlocking procedures for locking and unlocking the shared resource by a successful lock requestor, and a lock override procedure responsive to the detection of the predetermined indication of protocol failure;

providing for the shared resource, an associated main lock data structure stored in a shared memory accessible by the plurality of processors, the main lock data structure including in a single atomic structure, the resources needed to lock the shared resource by a successful lock requestor, to establish a queue of unsuccessful lock requestors, and to validate the existence of the lock;

providing for the shared resource, an associated auxiliary lock data structure stored in a shared memory accessible by the plurality of processors, the auxiliary lock data structure including the resources needed to identify the successful lock requestor'"'"'s place in a queue of requestors and to identify the successful lock requestor;

detecting, by one of the processors, a predetermined indication of protocol failure and identifying the failing processor by the detecting processor;

initiating the lock override procedure by the detecting processor responsive to the predetermined indication of protocol failure; and

,

in a single atomic operation by the detecting processor, examining the contents of the auxiliary lock data structure to determine if the identified failing processor is the successful lock requestor, and either, if the identified failing processor is the successful lock requestor, in a single atomic operation by the detecting processor, examining the contents of the main lock data structure and writing data to the main lock data structure to reserve the lock to the next requestor in the queue of unsuccessful lock requestors and to revalidate the lock, or, if the identified failing processor is not the successful lock requestor, exiting the cooperative lock override procedure.

2. A method according to claim 1 wherein the lock services procedure further includes at least two lock mode procedures and a lock mode selection procedure for selecting one from the lock mode procedures by a successful lock requestor, wherein the locking and unlocking procedures include one or more procedures for locking and unlocking the shared resource in the selected lock mode by a successful lock requestor, wherein the cooperative lock override procedure is selectively associated with a lock mode, wherein the atomic main lock data structure further includes the resources needed to identify one of the lock modes, and wherein the auxiliary lock data structure further includes the resources needed to identify one from the lock modes.

3. A method according to claim 1 wherein the method further includes, prior to examining the contents of the main lock data structure by the detecting processor, the step of in a single atomic operation by one of the requesting processors, examining the contents of the main lock data structure to determine if another requesting processor has previously locked the shared resource and if the lock contents are valid, determining that the contents are invalid or no other requesting processor has previously locked the shared resource, and writing data to the main lock data structure to reserve and validate the lock.

4. A method according to claim 2 wherein, the step of examining the contents of the main lock data structure by the detecting processor includes the step of, in the same atomic operation, verifying by examining the contents of the main lock data structure that the identified lock mode is the lock mode associated with the first lock override procedure and the step of writing data to the main lock data structure to reserve the lock to the next requestor in the queue of unsuccessful lock requestors and to revalidate the lock, includes the step of, in the same atomic operation, the step of invalidating the identified lock mode.

5. An intelligent data storage system comprising:

a shared resource;

a plurality of processors as lock requestors, each processor supporting atomic operations and being coupled to the shared resource through one or more first common communication channels, and each processor implementing a monitoring procedure for detecting a predetermined indication of protocol failure by any one of the plurality of processors and identifying the failing processor;

a shared memory accessible over one or more second common communications channels to all of the processors;

a lock services procedure implemented in each of the processors, the lock services procedure including a queuing procedure for unsuccessful lock requestors, and locking and unlocking procedures for locking and unlocking the shared resource by a successful lock requestor, and a cooperative lock override procedure responsive to the detection of the predetermined indication of protocol failure;

an atomic main lock data structure, responsive to the lock services procedures, implemented in the shared memory and associated with the shared resource, which includes the resources needed to lock a shared resource by a successful lock requestor, to establish a place in a queue of unsuccessful lock requestors, and validate the existence of the lock;

an atomic auxiliary lock data structure, responsive to the lock services procedures, implemented in the shared memory and associated with the shared resource, which includes the resources needed to identify the successful lock requestor'"'"'s place in a queue of requestors and to identify the successful lock requestors;

each processor other than the successful lock requestor being operable in accordance with its monitoring procedure to detect a predetermined indication of protocol failure and identify the failing processor; and

in accordance with its lock services procedure, first to initiate its cooperative lock override procedure responsive to its detection of the predetermined indication of protocol failure, and then in a single atomic operation, to examine the contents of the auxiliary lock data structure to determine if the identified failing processor is the successful lock requestor, and either, if the identified failing processor is the successful lock requestor, in a single atomic operation, to examine the contents of the main lock data structure and write data to the main lock data structure to reserve the lock to the next requestor in the queue of unsuccessful lock requestors and to revalidate the lock, or, if the identified failing processor is not the successful lock requestor, to exit the cooperative lock override procedure.

6. An intelligent data storage system according to claim 5 wherein the lock services procedure further includes at least two lock mode procedures, a lock mode selection procedure for selecting one from the lock mode procedures by a successful lock requestor, wherein the locking and unlocking procedures include one or more procedures for locking and unlocking the shared resource in the selected lock mode by a successful lock requestor, wherein the cooperative lock override procedure is selectively associated with a lock mode, and wherein the atomic main lock data structure includes the resources needed to identify one of the lock modes.

7. A method according to claim 6 wherein each requesting processor is further operable in accordance with its lock services procedure to verify in examining the contents of the main lock data structure in an atomic operation that the identified lock mode is the lock mode associated with the cooperative lock override procedure and to invalidate the identified lock mode.

8. An intelligent data storage system according to claim 5 wherein each requesting processor is further operable in accordance with its lock services procedure, in a single atomic operation, to examine the contents of the main lock data structure to determine if another requesting processor has previously locked the shared resource and if the lock contents are valid, and either, if the lock contents are valid and some other requesting processor has previously locked the shared resource, to write data to the main lock data structure to establish its place in a queue of requestors for subsequent locks on the shared resource, or if the contents are invalid or no other requesting processor has previously locked the shared resource, to write data to the main lock data structure to reserve and validate the lock.